SAS takes on Social Media Analytics

Enterprise business intelligence analysts are getting a new toy to play with, this one addressing the rapidly evolving field of social media analytics. SAS, practically our neighbor just up the road in Cary, NC, launched their new Social Media Analytics solution this past week in Seattle – a launch which incidentally made some very smart uses of social media in its own right. As a follow up, SAS hosted a small event Wednesday night for agencies and social media practitioners in the Research Triangle area, and was nice enough to throw us an invite or two.

A quick thanks to David Thomas, the social media manager from SAS, for both those invites and for hosting a great event. You can find some photos of the gathering over on Wayne Sutton’s blog. Beyond the networking, David managed to work in a very interesting demo of the new service. Here are two of my initial takeaways:

Making Sentiment Analysis Useful

Measuring the volume of social mentions around your brand is extremely useful, but understanding what they actually mean is what often matters most. This is where sentiment analysis comes into play. It’s an effort to apply automated scoring (generally Positive/Negative/Neutral) via Natural Language Processing to large numbers of tweets, blog posts, forum threads, and status updates, in order to gain an understanding on the general tone of the social conversation. Many vendors offer it, including Radian6 and Sysomos, both of which we regularly use here at Ignite, and it has become something of a must-have feature for any current social media analytics engine.

Sentiment analysis however very often treads a fine line between science and intuition, and it’s our view that you need to take it with at least a moderately sized grain of salt. Our own Brian Friedlander took a closer look at social media sentiment analysis a few months ago, and came away with a catiously optimistic wait-and-see outlook.

All of this is relevant because SAS made a significant claim a couple times during the launch event. Based on a follow up analysis (actual humans validating the scores their analytics engine assigned), SAS claims to have achieved a 90% accuracy rate in their sentiment scoring. This will be a very impressive level if it proves out over time and across real world customer implementations, something I’m hoping we’ll see covered in future reports from SAS. It’s also one feature that could be a game changer for SAS, as highly accurate sentiment scoring could turn the metric from a useful-but-vague indicator to a truly practical tool for marketers at all levels.

Aiming for the Enterprise

SAS as a company tends to work at the enterprise level, and this new offering is clearly designed for large customers both in its functionality and in the price point, which places it out of reach for most small to medium business and agencies. What’s significant about this is it is clearly not targeting the frontline social media manager, but rather the enterprise marketing analytics teams that already use deep business intelligence tools in their day-to-day jobs. Where the social media manager might have a moderate budget and a heavy focus on day-to-day monitoring (plus an aversion to spending large chunks of their budget on anything else beyond execution), the marketing analytics groups within large organizations may be much more receptive to what SAS has to offer. After all, it’s basically just an extension of the tools they’re already comfortable with.

SAS seems to be betting on taking social media analytics deep into the research back office, a sign that they feel the discipline has gained a much higher level of importance and visibility within large enterprises. We tend to agree with that – the calls for deeper and more insightful research based on richer data are growing louder, and it’s reasonable to see traditional-minded marketing analytics teams start to take more ownership of social media data and analysis. It’s a big bet from SAS, but the winds seem to be blowing in their favor.

David and team did indicate they plan on rolling out templates to simplify the setup process, which should help make the suite more accessible to both agencies and small and medium businesses, though no timeframe was given.

A parting note…

Social media analytics is a crowded field, with dozens of vendors of all sizes active, constantly delivering new products, and continually tweaking their features to support the latest and greatest social media tools and trends out there (analyzing location-based checkins on Foursquare or Gowalla, for example). For a fairly comprehensive view of what’s out there, in case your weekend reading load is looking a bit light, the team at Altimeter Group just released a report on the topic that’s worth the download.

Disclosure: We received an invite to this event on the SAS campus, a free beer or two, and some very tasty snacks, but no other compensation or consideration.

Ignite Social Media